Abstract
Canonical correlation analysis (CCA) is explored as a multivariate linear statistical methodology with which to forecast fluctuations of the El Niño/Southern Oscillation (ENSO) in real time. CCA is capable of identifying critical sequence of predictor patterns that tend to evolve into subsequent patterns that can be used to form a forecast.
The CCA model is used to forecast the 3-month mean sea surface temperature (SST) in several regions of the tropical Pacific and Indian oceans for projection times of 0 to 4 seasons beyond the immediately forthcoming season. The predictor variables, representing the climate situation in the four consecutive 3-month periods ending at the time of the forecast, are 1) quasi-global seasonal mean sea level pressure (SLP) and 2) SST in the predictand regions themselves. Forecast skill is estimated using cross-validation, and persistence is used as the primary skill control measure.
Results indicate that a large region in the eastern equatorial Pacific (120°−170°W longitude) has the highest overall predictability, with excellent skill realized for winter forecasts made at the end of summer. CCA outperforms persistence in this region under most conditions, and does noticeably better with the SST included as a predictor in addition to the SLP.
It is demonstrated that better forecast performance at the longer lead times would be obtained if some significantly earlier (i.e., up to 4 years) predictor data were included, because the ability to predict the lower-frequency ENSO phase changes would increase. The good performance of the current system at shorter lead times appears to be based largely on the ability to predict ENSO evolution for events already in progress.
The forecasting of the eastern tropical Pacific SST using CCA is now done routinely on a monthly basis for a 0-, 1-, and 2-season lead at the Climate Analysis Center. Further refinements, and expected associated increases in skill, are planned for the coming several years.